The third movement of the E minor Shostakovich piano trio begins with the pianist playing a series of whole note chords, all by their lonesome, before the violin finally comes in at the end of bar eight (click here to take a quick listen).
The tricky thing, is that each of those whole notes is held for quite a long period of time. Like anywhere from 4 to 6 seconds per note or longer, depending on what tempo the group chooses to take. Which is a lot of time for your mind to play tricks on you and make you wonder exactly where to place that next note.
Or at least, that’s what I remember the pianist I played this with one summer stressing about before a performance…
In any case, that’s the first thing that came to mind when I was talking with a pianist recently who noted that the legendary pianist Artur Schnabel once said something to the effect that “When you play fast, you should think slow, and when you play slow, you should think fast.”
In essence, that when you have a lot of fast notes, it helps to think about the longer arc of the phrase, and when playing long notes, it helps to fill them in by thinking in terms of smaller subdivisions or pulses. And not for metronomic precision of course, but to facilitate playing in a more musically compelling way.
Indeed, this is something that is evident in Leon Fleisher’s teaching as well (who studied with Schnabel). You can hear him speak to this a bit in this master class from 2015 (starting at about 50:50 through about 1:05:00).
Countless other musicians have spoken about the value of subdividing too – but does it actually work? Like, if we have long notes or rests to deal with, and there’s no conductor or other player to cue us in or keep us from distorting time without realizing it, is there any research supporting the use of subdividing as an effective tool for maintaining more accurate time and rhythm?
Oh, and if you were wondering what the deal is with my recent fascination with rhythm, well, in researching last week’s article, I kind of got sucked down a whole trail of other studies on rhythm… Turns out there’s actually quite a bit of intriguing research out there on rhythm amongst musicians. Don’t worry – I’m sure it’s just a phase that will pass soon enough. =)
But in the meantime, let’s take a look at some research on subdividing!
A subdividing study
A Yale researcher (Repp, 2010) recruited six graduate students from the School of Music – two pianists, a harpist, and a mix of string players.
All were asked to complete a series of four weekly test sessions, in which they were presented with forty 30-beat sequences. Where essentially, each sequence was like having seven bars of 4/4, with two quarter notes to set the tempo and lead you into bar one.
The beats in each sequence were held constant, and maintained a set tempo. But the researcher did vary the tempo of the sequences from one to the next. Where the time between beats ranged anywhere from 1 second to 3.25 seconds. In other words, some sequences were faster, and some were slower, as the researcher was curious to see if subdividing might be more useful, the slower the tempo. Which seems totally in line with Schnabel’s idea of thinking fast when playing slow, whether the researcher knew about this or not.
Four strategies for keeping time
Anyhow, each week, the musicians tried keeping consistent time using a different strategy.
One week, they were asked to tap in sync with each beat. They were also explicitly instructed to avoid subdividing in any way – “DO NOT SUBDIVIDE intervals in any way (either mentally or by moving).” (On-the-beat tapping)
Another week, they were asked to tap in sync with each beat, while mentally imagining a sound or movement on the offbeats – but without actually moving or making any sort of sound. (On-the-beat tapping with mental subdivision)
In another one of their sessions, they were asked to tap in sync with the downbeat, and also the offbeat. (Double tapping)
And on yet another week, they were asked to remain still on the downbeats, and tap only on the offbeats. (Offbeat tapping)
Then, the researcher measured how far off their taps were from the actual tones, and whether each of the taps were early or late (ahead of or behind the beat).
And did subdividing lead to more accurate tapping?
Did subdividing help?
Well, as you are undoubtedly not at all surprised to hear, yes. Subdividing did indeed lead to more accurate rhythm.
But let’s take a closer look, as some of the details are kind of interesting.
One thing that I thought was intriguing is that overall, across all tempos, participants’ taps tended to come before the beat. As in, there was a tendency to come in too soon, rather than too late. And apparently this is consistent with previous research too.
Which made me think of something Fleisher once said. My brain is a little fuzzy on the specific wording, but the gist was that we should place notes as late as possible – without actually being late.
So taken together, I guess maybe this means that our natural tendency is to compress time? I’d be curious to hear if that’s consistent with your own experience…
Further evidence that moving your body helps
Anyhow, before I go off on too much of a tangent, yes, the main finding was that whether subdividing mentally without any sound or movement, or physically tapping on the offbeats, subdividing did lead to taps that were significantly closer to the beat. That is to say, more accurate timing.
And for what it’s worth, being allowed to subdivide the offbeats while moving their body did seem to be substantially more effective than having to remain completely still and only subdividing mentally (which is consistent with the study we looked at last week).
But either way, the data suggests that subdividing is indeed an effective way to ensure we don’t compress time and clip notes or rests short.
And…when might subdividing be most useful?
One other interesting finding was that the benefits of subdividing became more and more apparent, the slower the tempo became.
In other words, when the tempo was faster, and there was just 1 second between beats, the difference in accuracy between subdividing and not subdividing was pretty small. But as the tempo slowed down, and there was in some cases as much as 3.25 seconds between the beats, the difference in accuracy between subdividing and not subdividing was really clear.
Based on the results, the researcher estimated that when beats are .9 seconds apart or less, subdividing might not lead to that much of a noticeable benefit in timing accuracy.
I’m still inclined to think there can be some benefits to subdividing even then, but as Schnabel suggested, there probably is a point at which the space between notes is so small that it makes more sense to switch over to thinking in terms of longer pulses rather than continuing to subdivide things into smaller and smaller pulses.
Ultimately, the goal with subdividing isn’t to develop more metronomic precision, of course, but to develop a stronger internal sense of pulse and rhythm.
If you haven’t explored this sort of thing in a systematic way before, a good place to start might be to play a piece you know pretty well – musically, of course – but while bisecting each beat as the participants in the study did. And maybe even trisecting or quadrisecting (yup, that’s a word – I looked it up!) some beats as Fleisher does in the video above (click here to see the exact section)
Take it a step further
Have you done a lot of metronome practice, but still feel like your internal sense of time isn’t reliable?
I figured this might be a good time to link to a couple videos, where two musicians draw on their backgrounds in neuroscience and kinesiology/motor learning and a bit of research to explain why the conventional way of practicing with a metronome can lead to it becoming a crutch, rather than a tool for helping us develop a better sense of rhythm (and what to do instead).
And one last thing…
If you’re an educator or teacher, and have been looking to add a few more strategies to your toolbox to help your students beat nerves, get into the zone, and enjoy more positive performance experiences (whether live or recorded), you might be interested in the live (online), 5-week workshop series – Performance Psychology Essentials for Educators – coming up in February.
Registration doesn’t begin until next week, but it’s specifically for teachers, and you can learn more about it here.
Or, if you’re already thinking you might sign up, please let me know what times would work best for you by filling out the quick survey here!
Repp, B. H. (2010). Self-Generated Interval Subdivision Reduces Variability of Synchronization with a Very Slow Metronome. Music Perception, 27(5), 389–397. https://doi.org/10.1525/mp.2010.27.5.389
- Who drew from his background in kinesiology/motor learning to explain the benefits of practicing at-tempo relative to slow practice in this podcast episode.
- Who drew on her background in neuroscience to explain how to memorize music more effectively in this podcast episode.